[SAC] errors with Trac

[Daniel Morissette]

Martin Spott wrote:
> 
> Already did that for a while today and noticed quite a few Apache and
> Postfix processes (load between 20 and 24) - nothing which made me feel
> surprised. To my opinion the web service checker is a bit too sensitive
> and reports unavailability when the system is simply responding a bit
> more slowly than usual.
> 

That's quite possible, it's just a simple script that uses "wget -T 5 
..." and if it fails the first time (i.e. no response in 5 seconds) then 
it waits 30 seconds and tries again. If it fails both times then it 
sends the error notice.

I have modified the script to increase the timeout from 5 to 15 seconds 
(i.e. "wget -T 15 ..."), let's see if that helps. Note that this just a 
legacy script that we had, and we use better alternatives to monitor 
servers these days such as Nagios, so I could disable the script if it 
creates too much noise.

Daniel
-- 
Daniel Morissette
http://www.mapgears.com/